Trump administration’s insistence on citizenship question may skew US Census by 9 million people

Meanwhile, a federal judge says new evidence suggests the Trump administration might have had discriminatory intent in seeking to add the question.

census document form and ball point ink pen on American flag for 2020
census document form and ball point ink pen on American flag for 2020

An estimated nine million Americans could go uncounted in the 2020 Census if the Trump administration is allowed to insert a citizenship question into the decennial headcount, according to the federal agency tasked with carrying out the survey.

A new U.S. Census Bureau research paper found that 8% of the population could be undercounted if the citizenship question is added, because households with noncitizen members might question the confidentiality of their responses and choose not to respond, fearing possible government misuse of the data.

According to the study, the Census Bureau will not have enough time to conduct a controlled trial that could accurately indicate the impact of the question on its head count of all people living in the country.

Adding the question, the bureau concludes, “could affect the self-response rate, a key driver of the cost and quality of a census.”


Meanwhile, the U.S. Supreme Court is weighing whether adding the question is constitutional, a decision that is expected any day.

And a lower court judge on Monday said new evidence shows the Trump administration may have added the question with discriminatory intent.

New documents uncovered from the files of now deceased GOP gerrymandering mastermind Thomas Hofeller show Commerce Secretary Wilbur Ross and the Trump administration lied that it wanted to add the citizenship question to enforce the Voting Rights Act.

The files, obtained by Common Cause, show Hofeller had previously urged the Trump administration to add the question to the Census, which would “clearly be a disadvantage to the Democrats” and “advantageous to Republicans and Non-Hispanic Whites.”

New emails provided to the court show Hofeller in 2015 discussed the citizenship question with Christa Jones, who is now a high-ranking Census Bureau official.


The “new evidence potentially connects the dots between a discriminatory purpose—diluting Hispanics’ political power—and Secretary Ross’s decision,” U.S. District Court of Maryland Judge George Hazel wrote in a new filing on Monday.

“The evidence suggests that Dr. Hofeller was motivated to recommend the addition of a citizenship question to the 2020 Census to advantage Republicans by diminishing Hispanics’ political power,” Hazel added.

If the question is allowed to be included on the survey, Census Bureau data shows that the question could have a much larger impact on the survey’s accuracy than expected. This would most certainly affect the government’s ability to accurately fulfill its constitutional requirement of counting all people living in the United States, regardless of whether they are citizens or in the country illegally.

It would also leave minority communities with large immigrant populations underrepresented in congress and would affect the share of federal dollars many cities and states receive for schools, roads, and other essentials.

The most recent U.S. Census Bureau study factored in more recent American Community Survey data than it had used during its last report released last year. According to the Bureau’s calculations, a total 8.7 million people would be affected.

The findings of the new study are much more damning than a Census Bureau calculation last year which estimated a 5.8% undercount.


The results are even more sobering than those in a study from by Urban Institute that estimated that more than four million people could be undercounted, including a disportionate number of black and Latino Americans.